Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[mifki]

I read that mifki struck some sort of agreement to not try to merge TwbT into dfhack because of how this mod works.

No, why. There are some changes need to be done on my side towards better interaction with dfhack and other plugins.
We've never discussed merging into dfhack though.

Btw, I need to check if for minor df updates these offsets can be found semi-automatically with byte search. Don't know why I haven't tried before.

[topis]

as i understand it, to get use of the dfhack automatic address scanning, one would need to define dfhack data structures for the map rendering functions and classes.

[Warmist]

as i understand it, to get use of the dfhack automatic address scanning, one would need to define dfhack data structures for the map rendering functions and classes.

Dfhack does not have "automatic address scanning". It has some tools/scripts/best practices/experienced hackers to quicker find addresses when new version appears but there is no hard requirement for automatic scripts (though it would help and even more if e.g. the original person that found that address moves on from dfhacking).

There is zero df functions (except for virtual ones) in dfhack because it's hard to keep track of their calling convention when full program optimization is on (which is on for df) and internal functions can be called with basically random calling convention. I think ragundo is trying to add one that would allow to export very detailed world maps like in legends mode but better. AFAIK he's done windows and linux already.

I personally am sceptical about adding functions into dfhack but only because i tried it once and was too lazy/unexperienced to have anything productive.

[topis]

that's what i mean, the scripts somewhat automate finding the offsets.

and if the functions that twbt patches had data structure definitions, the scripts would find the offsets for new df versions easier.
https://dfhack.readthedocs.org/en/stable/library/xml/SYNTAX.html

as i understand it, to get use of the dfhack automatic address scanning, one would need to define dfhack data structures for the map rendering functions and classes.

Dfhack does not have "automatic address scanning". It has some tools/scripts/best practices/experienced hackers to quicker find addresses when new version appears but there is no hard requirement for automatic scripts (though it would help and even more if e.g. the original person that found that address moves on from dfhacking).

There is zero df functions (except for virtual ones) in dfhack because it's hard to keep track of their calling convention when full program optimization is on (which is on for df) and internal functions can be called with basically random calling convention. I think ragundo is trying to add one that would allow to export very detailed world maps like in legends mode but better. AFAIK he's done windows and linux already.

I personally am sceptical about adding functions into dfhack but only because i tried it once and was too lazy/unexperienced to have anything productive.

[endoftheline]

that's what i mean, the scripts somewhat automate finding the offsets.

and if the functions that twbt patches had data structure definitions, the scripts would find the offsets for new df versions easier.
https://dfhack.readthedocs.org/en/stable/library/xml/SYNTAX.html

as i understand it, to get use of the dfhack automatic address scanning, one would need to define dfhack data structures for the map rendering functions and classes.

Dfhack does not have "automatic address scanning". It has some tools/scripts/best practices/experienced hackers to quicker find addresses when new version appears but there is no hard requirement for automatic scripts (though it would help and even more if e.g. the original person that found that address moves on from dfhacking).

There is zero df functions (except for virtual ones) in dfhack because it's hard to keep track of their calling convention when full program optimization is on (which is on for df) and internal functions can be called with basically random calling convention. I think ragundo is trying to add one that would allow to export very detailed world maps like in legends mode but better. AFAIK he's done windows and linux already.

I personally am sceptical about adding functions into dfhack but only because i tried it once and was too lazy/unexperienced to have anything productive.

Actually, the PATCHES file in TWBT describes a pretty OK procedure for finding the offsets, that works if automated too:

Spoiler (click to show/hide)

Code: [Select]

$ ./get_twbt_addresses.py 42.05 linux
processing release 42.05 linux
advmode_render: [0x9255fc8+0x8]
searching for p_render_lower_levels (this is slow)...
0x08e8c510
found jump at 0x08e8c510
----------------------------------------
#define A_RENDER_MAP 0x08b9b720
#define A_RENDER_UPDOWN 0x08901f00
#define NO_DISPLAY_PATCH
static patchdef p_dwarfmode_render = { 0x0840b502, 5 };
static patchdef p_advmode_render[] = { { 0x083b306d, 5+7+5 }, { 0x083b3121, 5+7+5 }, { 0x083b373c, 5+7+5 }, { 0x083b3b6a, 5+7+5 } };
static patchdef p_render_lower_levels = { 0x08e92a20, 13, true, { 0x36,0x8b,0x84,0x24,0x14,0x00,0x00,0x00, 0x3e,0xc6,0x00,0x00, 0xC3 } };

For Linux it reproduces the results for 40.24, 42.03-42.04 versions, and produces an OK looking patch for .05 too. I hope to get Win and OSX support working in a day or two and  the finished script.

I'd appreciate any extra tips about finding those addresses in a cleaner, more robust way than described in the PATCHES file.

[mifki]

That's interesting. What does your script use?

[mifki]

I've added/merged all offsets for .05, but didn't make a build because I'm not sure if the develop dfhack branch is ok to use.

[endoftheline]

That's interesting. What does your script use?

radare2 with the r2pipe Python API. It basically does all the steps described in the PATCHES.md file.

the patches compiled and seemed to not crash the game, although I only had time to look around and embark site and verify that zoom is working as intended. I need to learn how to install tilesets...

Some documentation about how to get TWBT installed properly would be nice to have. I can read the source and look at existing packs but its a massive pain, and I'm never sure I've done it the right way. As it is I just stuck a symlink in a develop branch dfhack's plugins dir, added it to plugins/CMakeLists.custom.txt, ran make install, set the rendering mode to TWBT, embarked a new fortress for a few minutes and called it a day.

Also is it normal that Hopper crashes a lot?

[mifki]

That's interesting. What does your script use?

radare2 with the r2pipe Python API. It basically does all the steps described in the PATCHES.md file.

the patches compiled and seemed to not crash the game, although I only had time to look around and embark site and verify that zoom is working as intended. I need to learn how to install tilesets...

Some documentation about how to get TWBT installed properly would be nice to have. I can read the source and look at existing packs but its a massive pain, and I'm never sure I've done it the right way. As it is I just stuck a symlink in a develop branch dfhack's plugins dir, added it to plugins/CMakeLists.custom.txt, ran make install, set the rendering mode to TWBT, embarked a new fortress for a few minutes and called it a day.

Also is it normal that Hopper crashes a lot?

I'd like to write Hopper scripts to find offsets eventually. I've heard about radare2 of course, but never bothered to compile it for OS X and try. Hopper doesn't crash for me, but I haven't used it on Linux much.

There's nothing to install actually, just the plugin file and shadows.png to data/art folder (the latter isn't strictly required for it to work).
As for compiling, I build it separately from dfhack with my Makefile - just set correct paths to df and dfhack and probably change gcc version, and "make inst" will do the job. CMakeLists.txt was contributed by someone else.

[fricy]

radare2 with the r2pipe Python API. It basically does all the steps described in the PATCHES.md file.

the patches compiled and seemed to not crash the game, although I only had time to look around and embark site and verify that zoom is working as intended. I need to learn how to install tilesets...

Thx for your contribution!
You can find some preconfigured 42.05 sets here, and optionally use PyLNP for installation.

edit:

I've added/merged all offsets for .05, but didn't make a build because I'm not sure if the develop dfhack branch is ok to use.

Loads ok so far.

[endoftheline]

...

Thx for your contribution!
You can find some preconfigured 42.05 sets here, and optionally use PyLNP for installation.

edit:

I've added/merged all offsets for .05, but didn't make a build because I'm not sure if the develop dfhack branch is ok to use.

Loads ok so far.

Thanks!

Also for anyone starting out like me: don't forget to add multilevel 6 or something like that into dfhack.init
This made me recheck the patch addresses like 3 times

[funkydwarf]

wow, wrong thread....nothing to see here

[PeridexisErrant]

ETA on a build that supports DFHack 42.05-alpha1 ?

[mifki]

ETA on a build that supports DFHack 42.05-alpha1 ?

Was waiting for the dfhack repo tag, will do once I get home.

[mifki]

done